文献类型：期刊文献

中文题名：图像灰度密度分布计算模型及肺结节良恶性分类

英文题名：Pulmonary nodule image grey density distribution feature extraction algorithm and adenocarcinoma benign/malignant classification

作者：Vanbang L E[1];朱煜[1];郑兵兵[1];杨达伟[2];任晓东[1];Thiminhchinh Ngo[3]

机构：[1]华东理工大学信息科学与工程学院,上海200237;[2]复旦大学附属中山医院,上海200032;[3]河内高科技研究中心,越南河内10000

年份：2020

卷号：37

期号：1

起止页码：296

中文期刊名：计算机应用研究

外文期刊名：Application Research of Computers

收录：CSTPCD;;北大核心:【北大核心2017】；CSCD:【CSCD_E2019_2020】；

基金：国家自然基金青年基金资助项目(81500078);复旦大学附属中山医院临床研究专项基金资助项目(2016ZSLC05,2016ZSCX02).

语种：中文

中文关键词：肺结节分类;密度分布特征;K-均值

外文关键词：lung nodule classification;density distribution feature;K-means

摘要：提出一种基于密度分布的特征评估算法,同时引入模式识别模型来评估该方法的效率。首先,从肺部肿瘤图像中随机提取像素块集,通过K-均值聚类算法将其分为10类,根据CT图像中肺结节像素值和聚类中心的关系,提取出10维特征向量,利用随机森林分类器进行模型训练,进而判断肺结节良恶性水平。在CT图像公开数据集LIDC-IDRI上的实验表明,分类平均精度达到0.9008。实验结果对比分析表明,提出的特征表达方法具有更优的分类效果和更高的鲁棒性。
Aimed at lung nodule benign/malignant classification,this paper proposed an effective grey scale density distribution feature extraction algorithm which combined with pattern recognition models to evaluate the classification system.The proposed feature extraction algorithm firstly collected a large number of blocks from lung tumor images and determined the distance matrix by calculating the relationships among the image blocks.Then,it used K-means clustering methods to classify the current image blocks and obtained 10 cluster centers.After that,it calculated the distribution density features by mapping CT value of nodule image pixels with the 10 cluster centers and extracted a 10-dimensional feature vector.Finally,the algorithm divided the extracted feature vectors into training and testing set to identify lung adenocarcinomas risk levels by random forest classification model.This paper evaluated the classification framework in LIDC-IDRI dataset,and the average accuracy reached to 0.9008.The proposed method outperforms the most recent techniques,and the experimental results show great robustness of the proposed method for different lung CT image datasets.